James Mickens, an associate professor of computer science at Harvard, gave a talk about the various flaws of web browsers and how we might go about making them more secure and robust:
To fix the browser, we need to radically restructure the browser architecture…so in particular I propose that we dramatically narrow the responsibilities of the browser itself and give more responsibilities to web pages. So in this new model the web developer defines her own rendering engine and her own script engine and her own markup parser.
The idea that web developers in the future might have this much control over the underlying technology of the web is completely mind boggling. James gives some good examples as to why we need this much control and how web pages that don’t send their own markup parser or rendering engine down the wire would then progressively fallback to whatever the browser provides.
What’s most interesting to me however are the questions that James asks, rather than his suggestions. Namely, what should be the relationship between a web page and a browser? How much control should developers have? What technologies do we need to make a functioning website?